wavelte packet reconstruction
7 vues (au cours des 30 derniers jours)
Afficher commentaires plus anciens
hi,
actually i decomposed the signal with wpdec() and measured each node using wpcoef now i have to adjust those signals before reconstruction.
anybody pls help me in which manner the signal can be adjusted before reconstruction to get the better signal.
thank u in advance
0 commentaires
Réponse acceptée
Wayne King
le 29 Oct 2011
Yes, you can use wpthcoef(), that does what you want.
Unless you have a very specific way you want to modify them beside hard or soft thresholding. If that is the case:
Read the terminal node coefficients with
cfs = read(T,'data');
Keep the approximation coefficients
tnodes = leaves(T);
app_coefs = read(T,'data',tnodes(1));
Change cfs the way you want. Let coefs be the changed coefficients.
Put the new coefficients in a tree.
Tnew = write(T,'data',coefs);
Place the approximation coefficients back.
Tnew = write(T,'data', tnodes(1),app_coefs);
0 commentaires
Plus de réponses (7)
Wayne King
le 22 Oct 2011
Hi, You can use wpthcoef or wpdencmp for denoising. Or you can simply modify the terminal nodes of your original tree, then create a new tree and use write to write your new coefficients to the terminal nodes.
0 commentaires
Wayne King
le 23 Oct 2011
Hi, It doesn't make sense in a wavelet packet tree to adjust each node separately because the higher levels nodes depend on the terminal nodes. It's not like a wavelet tree where you can modify just the detail coefficients.
You should decompose the signal down to whatever level you wish and just modify the terminal nodes. By definition, everything above it will be modified.
0 commentaires
thangam
le 23 Oct 2011
2 commentaires
Wayne King
le 23 Oct 2011
wavedec() does not give you a wavelet packet transform. That gives you a wavelet transform, so which did you do?
Voir également
Catégories
En savoir plus sur Signal Analysis dans Help Center et File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!